Detection of Melanoma Skin Cancer using Dermoscopic Skin Lesion Images

Melanoma is a serious type of cancer. This belongs to the skin cancer category and develops from the abnormal growth of melanocytes on our skin. A controlled growth of melanocytes is benign in nature and develops normal lesions. An uncontrolled growth of melanocytes is malignant in nature and develops abnormal lesions on the body. These abnormal lesions are cancerous in nature and this condition is known as melanoma. The benign and malignant lesions can be differentiated with their shape, texture and color features. In this paper, melanoma skin cancer detection is done by the classification of benign and malignant skin lesions from dermoscopic images. These dermoscopic images are preprocessed and then the required part of the image is obtained by segmentation. The statistical texture features and shape features are extracted from the segmented image. These extracted features are provided as input to several classifiers and the classification of benign and malignant skin lesions has been done. As a result, the melanoma skin cancer is detected from the dermoscopic image. Among the classifiers employed, Support Vector Machine (SVM) classifier with a linear kernel produces good classification results with efficient performance metrics.